The buccal micronucleus cytome (BMNcyt) assay in uncultured exfoliated epithelial cells from oral mucosa is widely applied in biomonitoring human exposures to genotoxic agents and is also proposed as a suitable test for prescreening and follow-up of precancerous oral lesions. The main limitation of the assay is the large variability observed in the baseline values of micronuclei (MNi) and other nuclear anomalies mainly related to different scoring criteria. The aim of this international collaborative study, involving laboratories with different level of experience, was to evaluate the inter- and intra-laboratory variations in the BMNcyt parameters, using recently implemented guidelines, in scoring cells from the same pooled samples obtained from healthy subjects (control group) and from cancer patients undergoing radiotherapy (treated group). The results indicate that all laboratories correctly discriminated samples from the two groups by a significant increase of micronucleus (MN) and nuclear bud (NBUD) frequencies and differentiated binucleated (BN) cells, associated with the exposure to ionizing radiation. The experience of the laboratories was shown to play an important role in the identification of the different cell types and nuclear anomalies. MN frequency in differentiated mononucleated (MONO) and BN cells showed the greatest consistency among the laboratories and low variability was also detected in the frequencies of MONO and BN cells. A larger variability was observed in classifying the different cell types, indicating the subjectivity in the interpretation of some of the scoring criteria while reproducibility of the results between scoring sessions was very good. An inter- laboratory calibration exercise is strongly recommended before starting studies with BMNcyt assay involving multiple research centers.